This conference explores the problems involved in producing measures of price change and quantity change. Historically the focus has been on measuring price changes using prices indices and using these to derive quantity changes. Price indices themselves are constructed from price data collected mainly from shops by combining them using weights derived from household or business surveys.
An important purpose behind the production of price indices is to make it possible to examine how welfare changes over time. This raises problems of its own. A particular issue can arise over the provision of publicly-provided goods which, even if not provided free, may not be sold at free-market prices. As a result, prices may not fully capture the value of these goods.
The increasing use of scanners in shops and the rising importance of web transactions offer new sources of data. In particular, scanner data make available timely quantity and price data, and raise the possibility of constructing price indices which reflect quantity responses in real time. At the same time, these new data sources raise new questions. How to construct real time price indices? How to cope with extremely large volumes of irregularly sampled data? How to address missing data problems in scanner data?
In addition, in both traditional and new sources of data, substantial issues remain in identifying quality changes, and thus producing satisfactory decompositions of value changes into price, quality, and quantity changes. The issue is perhaps particularly acute with measuring housing prices where a significant component of price variation is driven by size, quality and location. A number of additional important conceptual issues arise here such as how to measure the service flow form the stock of housing and how to properly account for risk and uncertainty.
Centre for microdata methods and practice